TADPOLE_BESS

1 TADPOLE and BESS

BESS Analysis

1.1 Loading the libraries

library("FRESA.CAD")
library(readxl)
library(igraph)
pander::panderOptions('digits', 3)
pander::panderOptions('table.split.table', 400)
pander::panderOptions('keep.trailing.zeros',TRUE)

1.2 Loading BSWiMS Results

opo <- par(no.readonly = TRUE)

load("~/GitHub/BSWiMS/TADPOLE_BSWIMS_Results.RData")
op <- opo

2 Predicting ADAS13

Here we will diagnose ADAS13

2.1 Learning ADAS13

BESSml <- BESS(ADAS13~.,TADPOLECrossMRITrain)

pander::pander(t(BESSml$selectedfeatures))

Table continues below

Gender

Hippocampus

ST128SV

ST1SV

ST2SV

ST68SV

M_ST24TA

M_ST25TA

M_ST44TA

M_ST54TA

M_ST57TA

M_ST62TA

M_ST13TS

M_ST15TS

M_ST23TS

M_ST25TS

M_ST39TS

M_ST44TS

M_ST48TS

M_ST54TS

M_ST56TS

M_ST14SA

M_ST15SA

M_ST23SA

M_ST24SA

M_ST31SA

M_ST129SA

M_ST35SA

M_ST39SA

M_ST47SA

M_ST50SA

M_ST51SA

M_ST55SA

M_ST57SA

M_ST40CV

M_ST45CV

M_ST46CV

Table continues below

M_ST52CV

M_ST60CV

M_ST17SV

M_ST30SV

M_ST61SV

M_ST65SV

RD_ST13TA

RD_ST15TA

RD_ST23TA

RD_ST31TA

RD_ST32TA

RD_ST35TA

RD_ST43TA

RD_ST44TA

RD_ST46TA

RD_ST49TA

RD_ST50TA

RD_ST55TA

RD_ST56TA

RD_ST60TA

RD_ST62TA

RD_ST13TS

RD_ST14TS

RD_ST24TS

RD_ST32TS

RD_ST40TS

RD_ST43TS

RD_ST46TS

RD_ST47TS

RD_ST48TS

RD_ST50TS

RD_ST57TS

RD_ST59TS

Table continues below

RD_ST60TS

RD_ST62TS

RD_ST15SA

RD_ST24SA

RD_ST32SA

RD_ST129SA

RD_ST35SA

RD_ST38SA

RD_ST44SA

RD_ST47SA

RD_ST50SA

RD_ST57SA

RD_ST58SA

RD_ST59SA

RD_ST60SA

RD_ST14CV

RD_ST15CV

RD_ST24CV

RD_ST25CV

RD_ST26CV

RD_ST32CV

RD_ST34CV

RD_ST35CV

RD_ST36CV

RD_ST40CV

RD_ST44CV

RD_ST46CV

RD_ST52CV

RD_ST54CV

RD_ST56CV

RD_ST62CV

RD_ST16SV

RD_ST18SV

RD_ST21SV

RD_ST61SV

RD_ST65SV

RD_ST66SV

prreg <- predictionStats_regression(cbind(TADPOLECrossMRITest$ADAS13,predict(BESSml,TADPOLECrossMRITest)),"ADAS13")

ADAS13

pander::pander(prreg)

• corci:

  cor
  0.661	0.617	0.701

• biasci: -0.4464, -0.9868 and 0.0941

• RMSEci: 7.32, 6.95 and 7.72

• spearmanci:

  50%	2.5%	97.5%
  0.62	0.571	0.665

• MAEci:

  50%	2.5%	97.5%
  5.72	5.39	6.07

• pearson:

  Pearson’s product-moment correlation: predictions[, 1] and predictions[, 2]

  Test statistic	df	P value	Alternative hypothesis	cor
  23.3	702	1.13e-89 * * *	two.sided	0.661

par(op)

2.2 Decorrelated

BESSmlD <- BESS(ADAS13~.,TADPOLECrossMRITrainD)

prreg <- predictionStats_regression(cbind(TADPOLECrossMRITestD$ADAS13,predict(BESSmlD,TADPOLECrossMRITestD)),"ADAS13")

ADAS13

pander::pander(prreg)

• corci:

  cor
  0.661	0.618	0.701

• biasci: -0.596, -1.136 and -0.056

• RMSEci: 7.32, 6.96 and 7.72

• spearmanci:

  50%	2.5%	97.5%
  0.648	0.599	0.691

• MAEci:

  50%	2.5%	97.5%
  5.58	5.23	5.93

• pearson:

  Pearson’s product-moment correlation: predictions[, 1] and predictions[, 2]

  Test statistic	df	P value	Alternative hypothesis	cor
  23.4	702	8.94e-90 * * *	two.sided	0.661

par(op)

3 Diagnosis MCI vs AD

3.1 Learning

#TADPOLE_DX_TRAIN$DX <- as.factor(TADPOLE_DX_TRAIN$DX)

BESSDXml <- BESS(DX~.,TADPOLE_DX_TRAIN)
pr <- predict(BESSDXml,TADPOLE_DX_TEST,type="response")

prBin <- predictionStats_binary(cbind(TADPOLE_DX_TEST$DX,pr),"MCI vs Dementia")

MCI vs Dementia

pander::pander(prBin$aucs)

est	lower	upper
0.624	0.567	0.681

pander::pander(prBin$accc)

est	lower	upper
0.687	0.638	0.733

pander::pander(prBin$berror)

50%	2.5%	97.5%
0.386	0.33	0.442

pander::pander(prBin$sensitivity)

est	lower	upper
0.469	0.368	0.573

par(op)

3.1.1 Decorrelated ML

BESSDXmlD <- BESS(DX~.,TADPOLE_DX_TRAIND)


prBin <- predictionStats_binary(cbind(TADPOLE_DX_TESTD$DX,predict(BESSDXmlD,TADPOLE_DX_TESTD,type="response")),"MCI vs Dementia")

MCI vs Dementia

pander::pander(prBin$aucs)

est	lower	upper
0.662	0.606	0.718

pander::pander(prBin$accc)

est	lower	upper
0.734	0.686	0.777

pander::pander(prBin$berror)

50%	2.5%	97.5%
0.34	0.282	0.394

pander::pander(prBin$sensitivity)

est	lower	upper
0.51	0.407	0.613

par(op)

4 Diagnosis NL vs AD

4.1 Learning

BESSDXmlNLDE <- BESS(DX~.,TADPOLE_DX_NLDE_TRAIN)

prBin <- predictionStats_binary(cbind(TADPOLE_DX_NLDE_TEST$DX,predict(BESSDXmlNLDE,TADPOLE_DX_NLDE_TEST,type="response")),"NL vs Dementia")

NL vs Dementia

pander::pander(prBin$aucs)

est	lower	upper
0.849	0.8	0.899

pander::pander(prBin$accc)

est	lower	upper
0.821	0.771	0.864

pander::pander(prBin$berror)

50%	2.5%	97.5%
0.198	0.148	0.252

pander::pander(prBin$sensitivity)

est	lower	upper
0.735	0.636	0.819

par(op)

4.2 Decorrelated Set

4.2.1 Learning

BESSDXmlNLDED <- BESS(DX~.,TADPOLE_DX_NLDE_TRAIND)

prBin <- predictionStats_binary(cbind(TADPOLE_DX_NLDE_TESTD$DX,predict(BESSDXmlNLDED,TADPOLE_DX_NLDE_TESTD,type="response")),"NL vs Dementia")

NL vs Dementia

pander::pander(prBin$aucs)

est	lower	upper
0.852	0.803	0.901

pander::pander(prBin$accc)

est	lower	upper
0.821	0.771	0.864

pander::pander(prBin$berror)

50%	2.5%	97.5%
0.212	0.168	0.267

pander::pander(prBin$sensitivity)

est	lower	upper
0.673	0.571	0.765

par(op)

5 Prognosis MCI to AD Conversion

Cox modeling of MCI to Dementia Conversion

5.1 Learning

Learning COX and ploting performance

bConvml <- BESS(Surv(TimeToEvent,status)~.,TADPOLE_Conv_TRAIN)

ptestl <- predict(bConvml,TADPOLE_Conv_TEST,type="response")
ptestr <- exp(ptestl)
boxplot(ptestl~TADPOLE_Conv_TEST$status)

boxplot(ptestr~TADPOLE_Conv_TEST$status)

perdsurv <- cbind(TADPOLE_Conv_TEST$TimeToEvent,
                  TADPOLE_Conv_TEST$status,
                  ptestl,
                  ptestr)

if (max(ptestl)>0 && min(ptestl)<0 )
{
  prSurv <- predictionStats_survival(perdsurv,"MCI to  AD Conversion")
  pander::pander(prSurv$CIRisk)
  pander::pander(prSurv$CILp)
  pander::pander(prSurv$spearmanCI)
}

5.2

50%	2.5%	97.5%
0.295	0.101	0.476

prBin <- predictionStats_binary(cbind(TADPOLE_Conv_TEST$status,ptestl),"MCI to  AD Conversion")

MCI to AD Conversion

pander::pander(prBin$aucs)

est	lower	upper
0.737	0.676	0.799

pander::pander(prBin$CM.analysis$tab)

	Outcome +	Outcome -	Total
Test +	69	55	124
Test -	29	120	149
Total	98	175	273

par(op)

5.3 Decorrelated

The decorrelated data set ### Learning Learning

bConvmlD <- BESS(Surv(TimeToEvent,status)~.,TADPOLE_Conv_TRAIND)

ptestl <- predict(bConvmlD,TADPOLE_Conv_TESTD)
ptestr <- exp(ptestl)
boxplot(ptestl~TADPOLE_Conv_TEST$status)

boxplot(ptestr~TADPOLE_Conv_TEST$status)

perdsurv <- cbind(TADPOLE_Conv_TEST$TimeToEvent,
                  TADPOLE_Conv_TEST$status,
                  ptestl,
                  ptestr)


if (max(ptestl)>0 && min(ptestl)<0 )
{
  prSurv <- predictionStats_survival(perdsurv,"MCI to  AD Conversion")
  pander::pander(prSurv$CIRisk)
  pander::pander(prSurv$CILp)
  pander::pander(prSurv$spearmanCI)
}

prBin <- predictionStats_binary(cbind(TADPOLE_Conv_TESTD$status,ptestl),"MCI to  AD Conversion")

MCI to AD Conversion

pander::pander(prBin$aucs)

est	lower	upper
0.825	0.774	0.875

pander::pander(prBin$CM.analysis$tab)

	Outcome +	Outcome -	Total
Test +	0	0	0
Test -	98	175	273
Total	98	175	273

par(op)

5.3.1 The End

.